INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

In fungi, many transcriptional regulators are characterized by the presence of a zinc finger containing a cluster of 6 cysteines (consensus sequence: CX₂CX₆CX_5-9CX₂CX_6-8C) that are bound by two zinc atoms and are referred to as zinc cluster proteins or Zn₂Cys₆ binuclear cluster proteins. Over 50 have been characterized in the yeast Saccharomyces cerevisiae (1, 2). The Zn₂Cys₆ domain is usually followed by a short linker region that bridges the zinc finger to the dimerization domain consisting of a coiled-coil consisting of heptad repeats (1). Zinc cluster proteins usually bind as homodimers to target DNA sequences.

Gal4p, which is required for activation of genes related to galactose metabolism, recognizes an inverted repeat consisting of CGG triplets spaced by 11 base pairs (bp)¹ (CGG N11 CCG), where N indicates nucleotides, with each zinc finger recognizing a CGG triplet (3). Some other members of this family bind to similar sites with an alternate spacing. For instance, Put3p, which activates genes related to proline metabolism, recognizes a DNA sequence related to that of Gal4p but with a spacing of 10 bp (CGG N10 CCG), whereas the activator of genes of the pyrimidine pathway, Ppr1p, binds to the sequence CGG N6 CCG (4-7). Determination of the solution or crystal structure of the DNA-binding domains of Gal4p, Ppr1p, and Put3p has revealed that their zinc fingers have a very similar structure which contacts the highly conserved CGG triplets (7-12) (reviewed in Ref. 13). Furthermore, zinc fingers of Gal4p, Ppr1p, and Put3p can be interchanged without altering the specificity of binding (14).

Other zinc cluster proteins also recognize CGG triplets but with a different orientation. For example, Hap1p, which activates genes required for cellular respiration, binds as a homodimer to a direct repeat consisting of two CGG triplets spaced by 6 bp (CGG N6 CGG) (15, 16). In addition, two other zinc cluster proteins, Leu3p and Pdr3p, bind to CGG triplets oriented in opposite directions, an everted repeat (CCG NX CGG) (17). Leu3p activates genes for the metabolism of branched amino acids and GDH1 which encodes NADP-dependent glutamate dehydrogenase, an enzyme involved in ammonium assimilation (18, 19).

Leu3p, which forms a homodimer in solution (20), is constitutively bound to target sites (21), and its activity is positively regulated by -isopropyl malate, an intermediate of leucine metabolism (22). Even though direct and inverted repeats consisting of CGG triplets are found in UAS_LEU3 (upstream activating sequence), we have shown, by mutational analysis, that Leu3p actually binds to an everted repeat containing two CGG triplets spaced by 4 bp (CCG N4 CGG). Alternatively, DNA targets of Leu3p could be considered to be an inverted repeat with the sequence CCG. However, a chimeric protein in which the zinc finger of Gal4p was replaced by the corresponding region of Leu3p was shown to recognize a Gal4p site (CGG N11 CCG) but not an inverted repeat with the sequence CCG N11 CGG (17). These results show that Leu3p does recognize the sequence CGG and strongly suggest that the two zinc fingers of Leu3p are oriented in opposite orientations to those of Gal4p (8).

Alteration of the spacing between the CGG triplets completely abolishes binding of Leu3p to a target site in vitro. Similarly, another member of this family, Pdr3p, a positive transcriptional regulator of multi-drug resistance genes (23, 24), binds to an everted repeat with no intervening nucleotides (CCGCGG) (17, 25, 26). Thus, diversity of binding sites is generated, in part, by three different DNA motifs (direct, inverted, and everted repeats) as well as differences between the number of nucleotides between the CGG triplets.

Other members of the family of zinc cluster proteins include Uga3p (also called Dur3p) and Dal81p (also called Uga35p or Durlp). Dal81p is a general positive regulator of genes involved in nitrogen utilization through catabolism of -aminobutyrate (GABA), urea, arginine, and allatoin (27, 28). The role of another member of the family of zinc cluster proteins, Uga3p, is restricted to genes related to GABA metabolism (29). Activity of Uga3p is dependent on the presence of GABA resulting in activation of the genes for GABA utilization as a nitrogen source such as the UGA1 and the UGA4 genes that encode GABA transaminase and GABA permease, respectively (28-31). Deletion of either UGA3 or DAL81 impairs activation of UGA1 and UGA4 (27-31). Unexpectedly, deletion of DNA sequences encoding the zinc finger of Dal81p does not affect its function suggesting that Dal81p may not act by binding to specific DNA sequences. Target sites have been identified in the UGA1 and UGA4 genes (31). These UASs contain three (or four) CGG triplets that could be recognized by Uga3p (see Fig. 1).

We have performed detailed mutagenesis of the UAS of UGA1. Our results show that Uga3p recognizes an everted CGG repeat spaced by 4 bp, a sequence similar to the target site of Leu3p. However, Leu3p does not recognize targets of Uga3p and vice versa. We show that additional specificity is provided by nucleotides located between the two CGG triplets.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Strains

Yeast (S. cerevisiae) strain YPH499 (32) (ura3-52 lys2-801 ade2-101 trp1-63 his3-2000 leu2-1) was used as wild type. Disruption of the UGA3 was done by transforming YPH499 with linearized plasmid pUGA-HIS (see below) and selecting for histidine auxotrophy. Deletion of DNA sequences encoding the zinc finger of LEU3 was performed by the method of Baudin et al. (33). The HIS3 gene was amplified by PCR using pHIS3 (see below) as template and the oligonucleotides primers 5'-GGAATAGAACTGGGATGAATGCTAGGAAAAGGAAATTCGCCTGCAGGGTTTTCCCAGTCA-3' and 5'-AAATTGATGATATTTTGTCGCAAATTCTTGAAACAGCTCAGCAGCGGATAACAATTTCAC-3'. Homologous recombination was verified by PCR or Southern blot analysis.

Plasmids

Disruption Plasmid and Expression Vector-- A portion of the UGA3 gene that encodes its DNA-binding domain (a.a 1-124) was PCR-amplified with Vent polymerase (New England Biolabs) using the primers 5'-CGGGATCCATGAATTATGGCGTGGAGAA-3' (initiator codon in bold) and 5'-GGAATTCAGTAGTTGTACAGCT-3' and yeast genomic DNA isolated from the strain YPH499 (32) as a template. The PCR product was cut by BamHI and EcoRI, purified on a column (PCR purification kit, Promega), and subcloned into pRSET-A (Invitrogen) or pGEX-f (17) cut with the same enzymes to give pUGA3-(1-124) and pGST-UGA3-(1-124), respectively. The HIS3 gene was amplified with the primers 5'-TTACTCTTGGCCTCCTCTAG-3' and 5'-GCCTCGTTCAGAATGACACG-3', and the PCR product was subcloned into the EcoRV site of pZERO-2 (Invitrogen) to give pMHIS3. The UGA3 disruption plasmid, pUGA-HIS, was constructed by subcloning the EagI fragment of pMHIS3 containing the HIS3 gene into the EagI site of pUGA3-(1-124). The resulting construct disrupts the UGA3 gene at nucleotide 244 relative to the ATG.

Reporters-- Reporter plasmids were constructed by inserting double-stranded oligonucleotides in front of a derivative of SLF178 containing a minimal CYC1 promoter driving lacZ transcription (34). Oligonucleotide 5'-TCGAAAAGCCGCGGGCGGGATTGTAC-3' and its complementary sequence 5'-AATCCCGCCCGCGGCTTT-3' were used to generate pUGA1-WT containing one copy of the UAS of UGA1 (31) in front of a minimal CYC1 promoter. Similarly, oligonucleotide 5'-TCGATCGGCCGGAACCGGCTTTGTAC-3' and its complementary sequence 5'-AAAGCCGGTTCCGGCCGA-3' were used to give pLEU2-WT containing one copy of the UAS of LEU2 (18). Reporter ULU, containing sequences corresponding to the UAS of LEU2 flanked by sequences of the UAS of UGA1, was constructed using the oligonucleotide 5'-TCGAAAAGCCGGAACCGGGATTGTAC-3' and its complementary sequence 5'-AATCCCGGTTCCGGCTTT-3'. Sequences of the mutants of the UASs of UGA1 and LEU2 are given in Tables I-III. All reporters were sequenced to verify the correct construction of the UASs.

-Galactosidase Assays

Strains were transformed with reporters and selected on plates lacking uracil. Colonies were grown overnight in YPD (35). Cells were then diluted in SD (35) supplemented with histidine, leucine, lysine, tryptophan, and adenine at 0.004% (w/v) and glucose (2% final concentration). Activity of Uga3p was induced by adding GABA to the medium (0.1% final concentration). -Galactosidase activity was measured with permeabilized cells (36). -Galactosidase assays were performed at least twice in duplicates.

In Vitro Binding of Leu3p and Uga3p

The DNA-binding domain of Leu3p (a.a. 1-147) was expressed in E. coli as a glutathione S-transferase fusion protein and purified as described (17). The glutathione S-transferase moiety was removed by cleavage with thrombin. The DNA-binding domain of Uga3p (a.a. 1-124) fused to a 6xHis tag was expressed as described (17) except that whole bacterial extracts were used for electrophoretic mobility shift assays (EMSA). EMSAs were performed as described (17). Competitor probes were generated by PCR using the primers 5'-AAGATGCGGCCAGCAAAACT-3' and 5'-CAGAGCACATGCATGCCATA-3' and the reporters as templates. The PCR products were then purified on a column (Qiagen). Labeled UAS of LEU2 (0.15 pmol) was added to 4, 25, 100, and 200 × excess of cold competitors, and Leu3 (a.a. 1-147) was added last. After 30 min at room temperature, the mixture was loaded on a 4% acrylamide gel. The percentage of binding (with no competitor taken as 100%) was measured (PhosphorImager, Fuji) and plotted against the log of the concentration of the cold competitor. Relative affinities were determined by comparing the concentration of each competitor required to give 50% binding as compared with the wild type competitor.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

The UASs UAS_LEU3 and UAS_UGA3 have 6 identical nucleotides (Fig. 1, Ref. 31). We first set out to determine if Leu3p and Uga3p have different target specificities and, if so, to identify which nucleotides are responsible for the discrimination between a Leu3p and a Uga3p target.

View larger version (50K): [in this window] [in a new window]   Fig. 1.   Top, comparison of the DNA sequences of the UASLEU3 (18, 19, 46, 47), the putative UAS of ILV5 (48), and the UASUGA3 (31). Conserved nucleotides found in nearly all the UASLEU3 or UASUGA3 are shown in bold with shadowed boxes. Bottom, CGG triplets critical for activity of the UAS of LEU2 or found in the UAS of UGA1 (labeled 1-3) are shown with arrows. Conserved nucleotides of the core sequence of the UASs of UGA1 and LEU2 are in bold.

The UAS of UGA1 and LEU2 Are Activated by Distinct Zinc Cluster Proteins-- We inserted the UAS of LEU2 or the UAS of UGA1 (or mutants) upstream of a minimal CYC1 promoter driving lacZ transcription. The reporters were transformed in a wild type strain or strains carrying disruption of genes encoding the transcriptional activators Uga3p or Leu3p, and their activity was assayed. As shown in Table I, activity of the UAS of UGA1 was induced 15-fold in a wild type strain grown in the presence of GABA and gave background levels in a uga3 strain. Deletion of the LEU3 gene had no effect on UGA1 activity indicating that Leu3p does not recognize the UAS of UGA1. As expected, activity of the UAS of LEU2 was reduced about 80 times in a strain carrying a deletion of the LEU3 gene, whereas the activity was reduced 3-4 times in a uga3 strain (Table I). If Uga3p was directly involved in the activation of LEU2, one would expect increased activity in the presence of GABA. However, the UAS of LEU2 showed slightly decreased activity in the presence of GABA in a wild type strain suggesting that the reduced activity is due to indirect effect. These results show that even if the UASs of UGA1 and LEU2 are related, they are activated by distinct zinc cluster proteins, Uga3p and Leu3p, respectively.

View this table: [in this window] [in a new window]   Table I In vivo activity of reporters containing one copy of the UAS of UGA1 or the UAS of LEU2 (or mutants), or no UAS, inserted in front of a minimal CYC1 promoter Reporters used in the assay and description of the UASs are given in the first and second rows, respectively. Core sequences of the UASs are given in the third row with mutations in bold. -Galactosidase of the reporters was measured in a wild type strain (WT) or strains carrying disruptions or deletions of the UGA3 or the LEU3 genes in the presence (+) or the absence () of GABA. For more details, see "Experimental Procedures."

Uga3p Recognizes an Everted Repeat Spaced by 4 bp-- If Uga3p recognizes CGG triplets, there are three triplets within the UAS of UGA1 that are potentially important for its activity (Fig. 1). To determine the importance of each of these CGG triplets, we mutated each one to CGA and assayed the activity of the resulting mutants. Mutating the first or the third CGG triplet of the UAS of UGA1 reduced -galactosidase activity 300- and 15-fold, respectively (Table I, mutants UGA1-A and UGA1-C). Similarly, mutations in both the first and third CGGs resulted in background activity (Table I, mutant UGA1-D). In contrast, a mutant bearing a nucleotide alteration in the second CGG triplet showed high activity that was dependent on Uga3p and GABA but not on Leu3p. Thus, CGG triplets numbers 1 and 3, which correspond to an everted repeat, appear to be essential for transcriptional activation by Uga3p.

In Vitro Binding of Uga3p to the UASs of UGA1 and Mutants-- We compared the activity in vivo of wild type or mutated UGA1 sequences with binding of the Uga3p DNA-binding domain to these sequences in vitro. With a probe corresponding to the UAS of UGA1, a retarded complex was formed in the presence of an extract prepared from cells containing the expression vector for Uga3p (a.a. 1-124) (Fig. 2) but not the parental vector (data not shown). Nucleotide alteration in the first CGG triplet (UGA1-A), but not the second (UGA1-B), abolished binding of Uga3p, an observation consistent with the relative function of these sequences in reporter plasmids in vivo. Mutation in the third CGG triplet (UGA1-C) resulted in a weak complex of identical mobility to that formed on the wild type sequence and a complex of faster mobility which may correspond to a monomer as observed for another zinc cluster protein (37). Mutations in both the first and the third CGG triplets (UGA1-D) completely abolished binding of Uga3p, an observation consistent with the background activity of the double mutant in vivo. Thus, the activity observed in vivo is correlated with in vitro binding of Uga3p.

View larger version (51K): [in this window] [in a new window]   Fig. 2.   Binding of Uga3p (a.a. 1-124) to the UAS of UGA1 and mutant sites. Sequences of the probes are given in the left panel. Mutations are shown in bold. EMSA is on the right panel with the probes indicated on top of the EMSA.

Nucleotides Located between the Everted CGGs Discriminate between a Uga3p and a Leu3p Site-- Our results suggested that nucleotides other than those present in the everted CGG motifs are responsible for discrimination between activation by Leu3p and Uga3p. To map these nucleotides, a reporter containing sequences encompassing the everted repeat of LEU2 flanked by sequences corresponding to the UAS of UGA1 was made (reporter ULU, Table II). The activity of that reporter was similar to the UAS of LEU2, high in a wild type strain even in the absence of GABA and reduced to background levels in a leu3 strain. These results suggest that key nucleotides for discrimination of a Leu3 and a Uga3p site are located between the two everted CGG triplets.

View this table: [in this window] [in a new window]   Table II In vivo activity of reporters containing one copy of the UAS of UGA1 or the UAS of LEU2 (or mutants), or no UAS, inserted in front of a minimal CYC1 promoter Reporters used in the assay and description of the UASs are given in the first and second rows, respectively. Core sequences of the UASs are given in the third row with mutations in bold. Sequence of the UAS of the reporter ULU is given under "Experimental Procedures." -Galactosidase of the reporters was measured in a wild type strain (WT) or strains carrying disruptions or deletions of the UGA3 or the LEU3 genes in the presence (+) or the absence of GABA (). NT, not tested. For more details, see "Experimental Procedures."

View this table: [in this window] [in a new window]   Table III In vivo activity of reporters containing one copy of the UAS of UGA1 or the UAS of LEU2 (or mutants), or no UAS, inserted in front of a minimal CYC1 promoter Reporters used in the assay and description of the UASs are given in the first and second rows, respectively. Core sequences of the UASs are given in the third row with mutations in bold. -Galactosidase of the reporters was measured in a wild type strain (WT) or strains carrying disruptions or deletions of the UGA3 or the LEU3 genes in the presence (+) or the absence of GABA (). For more details, see "Experimental Procedures."

Nucleotides in the Middle of the UAS of UGA1 Are Important for Uga3p Activity-- We determined the importance of the two central nucleotides of the UAS of UGA1. For the wild type UAS of UGA1, the sequence GGCGG is on the top strand of the everted repeat and the sequence CGCGG is on the lower strand (nucleotide difference in bold). Therefore, we tested if the presence of a G or a C at that position is important for activation by Uga3p. We constructed two reporters containing the perfect everted repeats GGCGG or CGCGG. Strong activity was observed with either reporter (UGA1-L, UGA1-M, Table III), giving rise to 25-fold induction in the presence of GABA. The results show that the "optimal" target sequence for Uga3p consists of an everted repeat with the sequence (G/C)GCGG.

Affinity of Leu3p to the UASs of LEU2, UGA1, and Mutants-- We correlated the in vivo activity of reporters described above with their ability to bind in vitro with the purified DNA-binding domain of Leu3p (a.a. 1-147; hereafter referred to as Leu3p). We measured the relative affinities of the UASs of LEU2 and UGA1 and mutants (Table IV). The affinity of UGA1 for Leu3p was weak (8%) compared with the affinity of the UAS of LEU2. The hybrid UAS consisting of the core sequence of LEU2 flanked by sequences of the UAS of UGA1 had similar affinity for Leu3p. In addition, the mutant of the LEU2 UAS, in which the central nucleotides were changed to those of the UGA1 UAS, had a reduced affinity for Leu3p (50%) relative to that of the wild type LEU2 UAS in agreement with its slightly reduced activity in vivo (4-fold). Thus, the Leu3p affinity for the mutant UASs is correlated with their in vivo activity.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

In this report, we show that Leu3p and Uga3p recognize highly related but distinct DNA sequences as follows: an everted CGG repeat spaced by 4 bp (CCG N4 CGG) and that nucleotides located between the CGG triplets are important for discrimination between a Uga3p and a Leu3p site. For example, any mutation targeting the two everted CGG triplets resulted in decreased activation by Uga3p (Table I), whereas mutating a third CGG triplet in the middle of the UAS had only minor effects in analogy to Leu3p (17). These results correlate with in vitro binding of the DNA-binding domain of Uga3p to the UAS of UGA1 and mutants (Fig. 2). Comparison with the UAS of UGA4 shows that only the two everted CGG triplets are conserved in both UGA1 and UGA4 (Fig. 1). Thus, Uga3p and Leu3p recognize a similar DNA motif, an everted sequence spaced by 4 bp (CCG N4 CGG). However, a target of Uga3p, the UAS of UGA1, was shown not to be activated by Leu3p. Conversely, the UAS of LEU2 is not a target of Uga3p.

We showed that nucleotides located between the two CGG triplets are responsible for discriminating between a Uga3p site and a Leu3p site. Changing either nucleotides that flank the two everted CGG triplets of the UAS of UGA1 to those found in the UAS of LEU2 abolishes activation by Uga3p (Table II, reporters UGA1-G and UGA1-H). Double nucleotide alteration results in very weak activity for both activators. However, a reporter (3XUGA1-I) containing three copies of the double mutant is responsive to Leu3p but not Uga3p. Thus the nucleotides that flank the two everted CGG triplets of the UAS of UGA1 are important but not sufficient for discrimination between Uga3p and Leu3p. In addition, a UAS containing the repeat TCGG or ACGG (reporters UGA1-J and UGA1-K, Table III) is almost inactive. Thus, there is a strict requirement for G or C nucleotides flanking the CGG motifs for activation by Uga3p and Leu3p, respectively.

Other nucleotides in the middle of the UAS of UGA1 are important for activity since a double mutation (GG to AA) showed low transcriptional activity (reporter UGA1-AA). Thus, our results suggest that the site recognized by Uga3p is an everted repeat with the repeat sequence of (G/C)GCGG. The two nucleotides in the middle of the UAS of LEU2 also play a role in modulating activity by Leu3p. For example, substitution of the two As found in the middle of the UAS of LEU2 to G results in decreased activation (4-fold) by Leu3p and in vitro binding. This is in agreement with the fact that most nucleotides found in the middle of the six known targets of Leu3p are A/T-rich and that none of them have only Gs or Cs at these positions (Fig. 1). We have previously shown (17) that single mutations (to A or T) outside the everted CGG triplets reduced in vivo activation and in vitro binding of Leu3p (17).

Three other members of the family of zinc cluster proteins bind to inverted CGG triplets with a spacing of 11 bp (Gal4p, Lac9p) or 6 bp (Ppr1p). Mutations in the CGG triplets have drastic effects on in vitro binding by these proteins, whereas they can tolerate multiple changes of the nucleotides located between the CGG triplets (5, 38, 39). This is consistent with crystallographic data of the DNA-binding domain of Gal4 and Ppr1p which showed that most of the contacts with DNA are made through interaction with the CGG triplets (8, 11). It is likely that Leu3p makes additional contacts with nucleotides between the CGG triplets allowing discrimination between the UAS of LEU2 and the UAS of UGA1, although this was not detected by methylation interference analysis (18, 20). Additional protein-DNA contacts would account for the strict DNA sequence requirement for Uga3p.

In addition to Uga3p, Dal81p is also required for induction of the GABA-responsive genes. The exact role of Dal81p in activation of UGA1 is not clear. Unexpectedly, deletion of the entire zinc finger of Dal81p had no effect on activation of UGA1 (28) similar to another zinc cluster protein (40). In addition, deletion of the DAL81 gene did not affect transcription of the UGA3 gene (31). Probably Dal81p does not contact DNA to form a heterodimeric complex as observed for the zinc cluster proteins Oaf1p and Oaf2p (41). One possibility is that Dal81p rather helps Uga3p to dimerize, resulting in increased affinity for target sites or provides an activation domain required for increased gene expression.

Our results may serve as a model for discrimination between UASs by other classes of zinc cluster proteins which recognize a specific CGG motif with a given spacing. For example, three zinc cluster proteins (Put3p, Tea1p, and Cha4p) have been shown to recognize the same DNA motif, inverted CGG triplets spaced by 10 bp (7, 42, 43). However, Tea1p does not activate Put3p sites (42). Some additional nucleotides may be responsible for the discrimination as we have observed in the present study for Leu3p and Uga3p.

In conclusion, alternate DNA motifs (direct, inverted, and everted repeats) and variation in the spacing between the CGG triplets have previously been shown to generate a diversity of binding sites for the family of zinc cluster proteins. In addition, other zinc cluster proteins do not recognize a pair of CGG triplets (44, 45) extending the repertoire of target sites for the family of zinc cluster proteins. Our results show that additional specificity can be provided by nucleotides flanking the CGG triplets generating distinct binding patterns of zinc cluster proteins to DNA motifs of a given spacing. This provides another mechanism by which the number of specific targets recognized by members of the family of zinc cluster proteins is increased.